Predicting the effect of feedstock on product yields and properties of theFCC process

The mechanism of petroleum refining processes are too complex, and no thoro ugh model has yet been developed. Neural networks represent an effective al ternative to mathematical modeling of refinery operations if a sufficient a mount of input-output data is available. In this paper, a feed forward neur al network that models the Fluid Catalytic Cracking (FCC) process will be p resented. The FCC process is the workhorse of the petroleum refining indust ry, making small and medium sized molecules out of big ones (gasoline and d istillate out of gas oils). The input-output data to the neural network was collected from the Literature on pilot and commercial plant operations and were obtained from actual refineries. Several network architectures were t ried and the network that best simulates the FCC process was retained. This network is able to predict yields of products of the FCC unit as well as t heir properties. The network consists of one hidden layer of twenty neurons , an input layer of four neurons, and an output layer of twelve neurons. Th e predictions of the neural network model were compared to those of a comme rcial simulator of the FCC process, to non-linear regression models, and to published charts. The results show that the neural network model consisten tly gives better predictions.